how long until Lexapro starts to work, likelihood Lexapro will work for your condition, and Lexapro vs. other Antidepressants

Approved & Off-label Uses | Lexapro Index | How to Take and Discontinue
Crazymeds Comprehensive Lexapro pages

1.  Comparatively Effective

Two of the most important things to know when deciding on which med is the best for a particular condition1: how likely is it to work and how long will it take.

The odds of a med working for a particular condition and how long it generally takes to work should be fairly easy to nail down, and not need to be summed up by the Internet shorthand YMMV (Your Mileage May Vary). Aside from it being hard enough to get an accurate diagnosis when brain cooties are involved, why is it so difficult to figure out if Lexapro is right for you and how long it will take for you to know that?

  • Because no one is quite sure exactly what causes various conditions.
  • Which is further complicated when everything is a spectrum disorder (e.g. bipolar 1, bipolar 2, all the others planned for DSM-V).
  • And they’re never really sure about how Lexapro works in the first place.
  • Plus, if you have more than one condition for which you’re taking one or more medications to treat, things get really complicated.
  • None of which is helped by studies that produce contradictory results, if they aren’t questionable in the first place.2

Always remember: if your symptoms suddenly get a lot worse, call your doctor immediately. Any drug that makes your symptoms worse is a drug you probably need to stop taking as soon as possible.

See our page on the tests researchers use to measure the efficacy of medications, including during clinical trials to get FDA approval.

2.  How Long Until Lexapro (escitalopram) Starts Working

Lexapro can start working within one week. As usual there are conflicting data. All the reports from people who’ve taken Lexapro that I’ve read (anecdotal evidence) indicate that Lexapro often starts working in the first week or two.

You should still give it at least three weeks, unless the side effects hit hard, fast, and don’t start getting better within two weeks of reducing the dosage (if there’s any room to reduce the dosage).

3.  How Effective Lexapro (escitalopram) is for its Approved Uses

3.1  Major Depressive Disorder (MDD)

The odds are pretty good that Lexapro will work for MDD, as in you have a 60–75% chance that you’ll respond to Lexapro3 - i.e. it will do something positive - and a 60–70% chance of remission, i.e. your MDD symptoms will more-or-less go away and stay away, although you may have to keep taking your meds to ensure you stay symptom-free.
Keep in mind that:

  1. Serotonin reuptake inhibition, and everything else SSRIs do, is not the answer for everyone, either by itself or at all.
  2. When you pool all the data together (see below) you’ll find that Celexa and Lexapro are better than the other SSRIs. It’s marginally statistically significant for everyone who still takes the other SSRIs. Their real advantage comes from sucking less than the others and being better at preventing relapses, so if you and your doctor think you’re going to be depressed forever without meds, you’re way better off with Celexa or Lexapro than another SSRI or other antidepressant.
  3. We can’t hammer this point home as hard or often enough: antidepressants work only for people who are seriously depressed.4

Like most SSRIs the clinical trials published in the PI sheet used a ridiculously small set of people with vague results. How small? The usually don’t print the number of patients, that’s how small. Things are different in Canada, where some actual numbers are published in the PI sheet. But it tells us only that people felt somewhat better, and not how many of them did.

Unlike most antidepressants I was actually able to find the clinical trials on PubMed, although there are no more data in the abstracts than in the PI sheet. Except for Fixed-dose trial of the single isomer SSRI escitalopram in depressed outpatients. The discrepancies between the PI sheet and the abstract that are really telling. The US PI sheet, which doesn’t allow you to copy any text, has the typically vague result of “statistically significant greater mean improvement compared to placebo on the MADRS.” But when you look at the study:

Clinical response was evaluated by the Montgomery-Asberg Depression Rating Scale (MADRS), the 24-item Hamilton Rating Scale for Depression (HAM-D), the Clinical Global Impressions (CGI) scales, the Hamilton Rating Scale for Anxiety (HAM-A), and patient-rated quality-of-life scales.
RESULTS: Escitalopram, at both doses, produced significant improvement at study endpoint relative to placebo on all measures of depression; significant separation of escitalopram from placebo was observed within I[sic] week of double-blind treatment. Citalopram treatment also significantly improved depressive symptomatology compared with placebo; however, escitalopram, 10 mg/day, was at least as effective as citalopram, 40 mg/day, at endpoint. Anxiety symptoms and quality of life were also significantly improved by escitalopram compared with placebo. The incidence of discontinuations due to adverse events for the escitalopram 10 mg/day group was not different from the placebo group (4.2% vs. 2.5%; p = .50), and not different for the escitalopram 20 mg/day group and the citalopram 40 mg/day group (10.4% vs. 8.8%; p = .83).

Note how more scales are mentioned in the abstract. The PI sheet omits two points that numerous studies will repeatedly prove: while Lexapro usually doesn’t work any better than Celexa, it often does work faster than Celexa, or any other SSRI.

The Australian PI sheet confirmed that the clinical trial about preventing the relapse of depression symptoms I found before is indeed the one in the PI sheet. It, too, is pitifully small (181 taking Lexapro, 93 taking placebo), but at least it had everyone taking their pills for nine months. Close to 75% of the people taking Lexapro after 36 weeks were still not depressed, vs. 60% taking the wonder drug Placebo.

You’re probably a lot less likely to commit suicide when taking Lexapro than taking a placebo. But is that really a big surprise? I guess it is to some people.

There have been studies experimenting with dosages up to 50mg(!) a day. This is the first time I’ve read a Lundbeck-sponsored study of Lexapro where there has been a significant number of people quitting due to side effects (20%). 30mg could be a dosage to consider, especially if you know or suspect you’re a rapid or ultra-rapid metabolizer of drugs cleared by CYP2D6 (substrates), or you take a drug that induces CYP2D6, like Tegretol (carbamazepine). 40mg seems to be the upper limit of real efficacy.

Predicting if a med will work for you is something that can be really useful. What have we got for Lexapro…

3.2  Generalized Anxiety Disorder (GAD)

The story for anxiety disorders is a bit better. People in the Panic/Anxiety community love Lexapro. Like MDD efficacy rate is still around 70–75%, but as GAD (and many of the other anxiety disorders for which it’s used off-label) tends to respond at lower dosages than MDD, the side effects tend to suck a lot less than GAD and other treatments.

In regards to GAD, this double-blind study of 158 people taking Lexapro vs. 157 people taking placebo had 68% of the people taking Lexapro responding against 41% taking the placebo. Not too bad for the placebo. In Lexapro’s favor, though, on the Hamilton Rating Scale for Anxiety, those taking Lexapro averaged a 11.3 point drop from a score of 18 points or higher, whereas those taking the placebo dropped only an average of 7.4 points. So you can fool some of the people with GAD into feeling better, but even those you fool won’t feel better than they could have with a real med.

What looks like a follow up to the clinical trials for Lexapro’s approval for GAD has even better numbers than the above. Those who completed the 8-week double-blind trials were given the option of knowing they were taking Lexapro (escitalopram oxalate) 56% of the people made it to the end, so about half in all, or roughly 250 people were still helped by Lexapro (escitalopram oxalate) for generalized anxiety disorder (GAD), and helped a lot.

Just like with MDD, Lexapro helps to prevent the relapse of generalized anxiety disorder. At least when compared with placebo. After taking Lexapro for 8–12 weeks:

4.  Likelihood Lexapro (escitalopram) will Work for Off-Label Applications

4.1  Panic-Anxiety Disorders other than GAD

Except for hoarding, where you’ll want Paxil, Lexapro may as well be approved for the rest of them. Oh, wait, Lexapro is approved to treat other anxiety disorders, just not in the US. We really need a way to streamline the approval process for meds when they’re already available here and approved for something else in practically every other country on the planet. In any event…

5.  Lexapro (escitalopram) versus Other Antidepressants for its Approved Indications

5.1  MDD

5.2  GAD

6.  How Lexapro (escitalopram) Compares with Other Drugs for Off-Label Treatments

  • Lexapro vs. Paxil for Social Anxiety Disorder (SAnD). This Lundbeck-sponsored study isn’t very fair, as it compares 20mg of the older, immediate-release Paxil against 5mg, 10mg, and 20mg of Lexapro. While 20mg of immediate-release Paxil is GSK’s recommendation for SAnD, it’s 25–37.5mg a day for the controlled-release flavor. While 10mg of Lexapro is equal to 10mg of Paxil, 20mg of Lexapro is more like 37.5mg of Paxil CR. So it’s not all that surprising that 20mg of Lexapro was more effective than 20mg of Paxil.
  • Lexapro vs. Celexa vs. Placebo for Panic disorders. Panic disorders are off-label for Celexa as well. The results - Lexapro and Celexa are both really good for panic, and Lexapro sucks less than the placebo. This sort of thing shows up in a lot of Lexapro studies and clinical trials: “The rate of discontinuation for adverse events was 6.3% for escitalopram, 8.4% for citalopram, and 7.6% for placebo.”
  • Lexapro vs. Celexa vs. Placebo for Panic disorders - the sequel. The results - Lexapro kicked Celexa’s ass, which is something you don’t see very often. This was done by DJ Stein, who is one of Lundbeck’s favorite researchers, so he could have set it up to slant things in Lexapro’s favor. The abstract isn’t very forthcoming in results: “Treatment with escitalopram was associated with significant improvement on all 5 subscales of the P&A. Citalopram was significantly different from placebo in 3 subscales. […] Escitalopram was superior to placebo on 12 of 16 items of the Q-LES-Q, while citalopram was superior on 7 items.” Unlike Dr. Stahl’s study above, there’s no hint in the abstract as to what “significant” means.

Keep Crazymeds on the air.
Donate some spare electronic currency
you have floating around The Cloud
Display your mad pride.    
Tell your imaginary friends about Crazymeds.

Approved & Off-label Uses | Lexapro Index | How to Take and Discontinue
Crazymeds Comprehensive Lexapro pages


1 Assuming you were correctly diagnosed in the first place.

2 Keep in mind that according to one study, most drug studies will skew in favor of the med made by the company that sponsored the study.* That's one of my favorite "no shit Sherlock" studies, although it did help in getting conflicts of interest showing up on papers.
Two additional papers along similar lines are Why Current Publication Practices May Distort Science** and Why Most Published Research Findings Are False***. So in addition to the books we use as source material, this is why we also factor a lot of anecdotal evidence (personal experience, experiences of people we know, case reports, what people have sent us in e-mail, and what is posted all over the Internet) into our conclusions regarding the likelihood of meds working, the prevalence of various side effects, etc.
While the drug companies are getting a lot more transparent and publishing more data in the PI sheets regarding the results of the clinical trials, they still don't publish how many times a drug failed a clinical trial.****

3 The data are often contradictory when compared with Celexa. Celexa averages 60-70% for both response and remission, while Lexapro is often in the neighborhood of 70-75%; hence the big spread on the response rate that seems a bit low given the results of a lot of studies, trials, and anecdotal evidence.

4 Most of the time at any rate. There's always an outlier study or two, as you'll see if you keep reading this page.

*Drug studies favoring sponsors the study.
**Why Current Publication Practices May Distort Science
***Why Most Published Research Findings Are False
****unpublished clinical trials

Date created {{$$newlycreated}} Page Author: Last edited by: JerodPoore on 2014–04–23

How Long Until Lexapro Works by JerodPoore is copyright {{$$yearly}} JerodPoore

Lexapro is a trademark of someone else. Look on the the PI sheet or ask Google who the owner is. The way pharmaceutical companies buy each other and swap products like Monopoly™ real estate, the ownership of the trademark may have changed without my noticing.

Page design and explanatory material by Jerod Poore, copyright © 2004 - 2014. All rights reserved.
Keep up with Crazymeds and and/or my slow descent into irreparable madness boring life. Pick your preferred social media target(s):

Jerod Poore
Wear my StraitjacketBatshit Crazy Blog
Crazymeds | Promote Your Page Too
Crazymeds on Facebook
Play Dress-Up with your Imaginary FriendsCrazymeds: The Blog

Almost all of the material on this site is by Jerod Poore and is copyright © 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, and 2014 Jerod Poore. Except, of course, the PI sheets - those are the property of the drug companies who developed the drugs the sheets are about - and any documents that are written by other people which may be posted to this site will remain the property of the original authors. You cannot reproduce this page or any other material on this site outside of the boundaries of fair use copying without the express permission of the copyright holder. That’s usually me, so just ask first. That means if want to print out a few pages to take to your doctor, therapist, counselor, support group, non-understanding family members or something like that - then that’s OK to just do. Go for it! Please. As long as you include this copyright notice and something along the lines of following disclaimer, I’m usually cool with it.

All rights reserved. No warranty is expressed or implied in this information. Consult one or more doctors and/or pharmacists before taking, or changing how you take any neurological and/or psychiatric medication. Your mileage may vary. What happened to us won’t necessarily happen to you. If you still have questions about a medication or condition that were not answered on any of the pages you read, please ask them on Crazy Talk: the Crazymeds Forum.
The information on Crazymeds pertains to and is intended for adults. While some information about children and adolescents is occasionally presented (e.g. US FDA approvals), pediatric-specific data such as dosages, side effects, off-label applications, etc. are rarely included in the articles on drugs or discussed on the forum. If you are looking for information regarding meds for children you’ll have to go somewhere else.
Know your sources!
Nobody on this site is a doctor, a therapist, or a pharmacist. We don’t portray them either here or on TV. Only doctors can diagnose and treat an illness. While it’s not as bad as it used to be, some doctors still get pissed off by patients who know too much about medications, so tread lightly when and where appropriate. Diagnosing yourself from a website is like defending yourself in court, you suddenly have a fool for a doctor. Don’t be a cyberchondriac, thinking you have every disease you see a website about, or that you’ll get every side effect from every medication1. Self-prescribing is as dangerous as buying meds from fraudulent online pharmacies that promise you medications without prescriptions.
All information on this site has been obtained through our personal experience and the experiences family, friends, what people have reported on various reputable sites all over teh intergoogles, the medications’ product information / summary of product characteristic (PI/SPC) sheets, and from sources that are referenced throughout the site. As such the information presented here is not intended as a substitute for real medical advice from your real doctor, just a compliment to it. You should never, ever, replace what a real doctor tells you with something from a website on the Internet. The farthest you should ever take it is getting a second opinion from another real doctor. Educate yourself - always read the PI/SPC sheet or patient information leaflet (PIL) that comes with your medications and never ever throw them away.
Crazymeds is not responsible for the content of sites we provide links to. We like them, or they’re paid advertisements, or they’re something else we think you should read to help you make an informed decision about a particular med. Sometimes they’re more than one of those things. But what’s on those sites is their business, not ours.
Very little information about visitors to this site is collected or saved. From time to time I look at search terms used and which pages they bring up in an effort to make the information I present more relevant. And the country of origin, just because I’m geeky like that. That’s about it. Depending on how you feel about Schrodinger, our privacy policy should either assuage or exacerbate your paranoia.
All brand names of the drugs listed in this site are the trademarks of the companies named on the PI/SPC sheet associated with the medication, sometimes on the pages about the drugs, even though those companies may have been acquired by other companies who may or may not be listed in this site by the time you read this. Or the rights to the drug were sold to another company. And any or all of the companies involved may have changed their names.
Crazymeds is optimized for the browser you’re not using on the platform you wish you had. Between you and me, it all looks a lot cleaner using Safari or Chrome, although more than half of the visitors to this site use either Safari or Internet Explorer, so I’m doing my best to make things look nice for IE as well. I’m using Firefox and running Windows 72. On a computer that sits on top of my desk. With a 23 inch monitor. Hey, at least you can make the text larger or smaller by clicking on the + or - buttons in the upper right hand corner. If you have Java enabled. Like 99% of the websites on the planet, Crazymeds is hosted on domain running an open source operating system with a variety of open source applications, including the software used to display what you’ve been reading. As such Crazymeds is not responsible for whatever weird shit your browser does or does not do when you read this site3.
No neurologists, psychiatrists, therapists or pharmacists were harmed in the production of this website. Use only as directed. Void where prohibited. Contains nuts. Certain restrictions may apply. All data are subject to availability. Not available on all mobile devices, in the 12 Galaxies Guiltied to a Zegnatronic Rocket Society, or in all dimensions of reality. Hail Xenu!

‘Everything is true, nothing is permitted.’ - Jerod Poore

1 While there are plenty of books to help you with hypochondria, for some reason there’s not much in the way of websites. Then again, staying off of the Internet is a large part of curing/managing the disorder.

2 Remember kids, Microsloth operating systems are like TOS Star Trek movies with in that every other one sucks way, way more. With TOS Star Trek movies you don’t want to bother watching the odd-numbered ones. With Microsloth OS you don’t want to buy and install the even-numbered ones. Anyone who remembers ME and Vista knows what I mean.

3 Have I mentioned how open source operating systems for commercial applications is one of the dumbest ideas in the history of dumb ideas?
[begin rant] I rent a dedicated server for Crazymeds. It’s sitting on a rack somewhere in Southern California along with a bunch of other servers that other people have rented. The hardware is identical, but no two machines have exactly the same operating systems. I don’t even need to see what is on any of the others to know this. If somebody got their server at the exact same time, with the exact same features as I did, I’m confident that there would be noticeable differences in some aspects of the operating systems. So what does this mean? For one thing it means that no two computers in the same office of a single company have the same operating system, and the techs can spend hours figuring out what the fuck the problem could be based on that alone. It also means that application software like IP board that runs the forum here has to have so many fucking user-configurable bells and whistles that even when I read the manual I can’t find every setting, or every location that every flag needs to be set in order for a feature to run the way I want it to run. And in the real world it means you can get an MBA not only with an emphasis on resource planning, but with an emphasis on using SAP - a piece of software so complex there are now college programs on how to use it. You might think, “But don’t people learn how to use Photoshop or Adobe Illustrator in college?” Sure, in order to create stuff. And in a way you’re creating stuff with SAP. But do you get a Bachelor of Fine Arts degree with an emphasis on Photoshop?
Back in the Big Iron Age the operating systems were proprietary, and every computer that took up an entire room with a raised floor and HVAC system, and had less storage and processing power than an iPhone, had the same operating system as every other one, give or take a release level. But when a company bought application software like SAP, they also got the source code, which was usually documented and written in a way to make it easy to modify the hell out of it. Why? Because accounting principles may be the same the world over, and tax laws the same across each country and state, but no two companies have the same format for their reports, invoices, purchase orders and so forth. Standards existed and were universally ignored. If something went wrong it went wrong the same way for everyone, and was easy to track down. People didn’t need to take a college course to learn how to use a piece of software.
I’m not against the open source concept entirely. Back then all the programmers read the same magazines, so we all had the same homebrew utilities. We even had a forerunner of QR Code to scan the longer source code. Software vendors and computer manufacturers sponsored conventions so we could, among other things, swap recipes for such add-ons and utilities. While those things would make our lives easier, they had nothing to do with critical functions of the operating system. Unless badly implemented they would rarely cause key application software to crash and burn. Whereas today, with open source everything, who the hell knows what could be responsible some part of a system failing. [/end rant]

Enable Crazymeds’ Financial Solvency!

Enable Crazymeds to keep spreading our knowledge. Donate some spare e-currency you have floating around The Cloud.

Improve Your Social Media Skills


Follow our Highly Irregular Updates and Paranoid Rants Other News


Show us teh like™

Square this Circle

Crazymeds for Twits

Sites That Probably Suck Less Than Crazymeds

Crazymeds Merchandise

Available at Straitjacket T-Shirts

Batshit Crazy shirts at Straitjacket T-Shirts

Clothes Hoist
T-Shirts, Hoodies & More $15-$43

Pile of Pills Mugs at Straitjacket T-Shirts

Get Mugged
11oz $13 & 15oz $14

Medicine is the Best Medicine Bumpersticker at Straitjacket T-Shirts

Stuck Up
All stickers $5

Mentally Interesting Button at Straitjacket T-Shirts

Button It!
2.25″ $4 & 3.5″ $4.50

Heart of Pills Earrings at Straitjacket T-Shirts

Dangly Bits
Earrings, Bracelets & More $15-$40

Pile Of Pills Gel Mousepad at Straitjacket T-Shirts

Yet more crap
$12 Mousepads to $85 Tablecloths

Burning Mind Books

Psychopharmacology tomes at Burning Mind Books

Books & Other Media for the Mentally Interesting

Replace my awful design choices with your own

Return to the Original Flavor